This methodological study aimed to identify the risk factors for reintubation within 72 hours following planned extubation in medical intensive care unit (ICU) patients and to develop and validate a prediction tool specialized for medical ICU patients...
This methodological study aimed to identify the risk factors for reintubation within 72 hours following planned extubation in medical intensive care unit (ICU) patients and to develop and validate a prediction tool specialized for medical ICU patients. The study included 609 adult medical ICU patients at a tertiary hospital in Korea who had undergone endotracheal intubation for more than 24 hours, successfully passed the ABC trials, and subsequently underwent planned extubation. Based on retrospective electronic medical record data (January 2022–December 2024), 305 patients were randomly assigned to the development group and 304 to the validation group using stratified randomization.
The study proceeded in five stages: (1) concept definition and item generation; (2) content validity evaluation by an expert panel (S-CVI = 0.95) and categorization ; (3) logistic regression analysis in the development group (n=305) to identify independent risk factors and develop the Prediction of ReIntubation in MEdical ICU patients (PRIME) score; (4) validity testing in the validation group (n=304) to assess sensitivity, specificity, positive/negative predictive values, and hypothesis testing; and (5) final score structure confirmation through Decision Curve Analysis (DCA). Multivariable logistic regression identified underweight (OR=2.34, p=.031), prolonged mechanical ventilation (≥10 days) (OR=2.76, p=.030), and hypoalbuminemia (<2.70 g/dL) (OR=2.04, p=.022) as significant predictors of reintubation. Considering clinical relevance, admission from a general ward (OR=1.84, p=.063), severe RASS (OR=2.58, p=.073), ABGA-pH acidosis (OR=2.72, p=.155), and ABGA-pH alkalosis (OR=1.91, p=.086) were also included in the final model. The PRIME score showed an AUROC of 0.71 (95% CI: 0.63–0.79) in the development group. In the validation group, using a cut-off score of ≥1 point, sensitivity was 60.6%, specificity 57.5%, positive predictive value 31.1%, and negative predictive value 82.2%. DCA demonstrated clinical usefulness within threshold probabilities of 0.15–0.30, with the optimal threshold of 0.25 aligning with the ROC-based cutoff point, confirming its appropriateness. The PRIME score provides a simple and explainable tool for objectively assessing the risk of reintubation before extubation in medical ICU patients. It can assist clinicians in early identification of high-risk patients and timely implementation of preventive interventions. Further multicenter prospective studies are needed to validate its external applicability and to establish post-extubation management protocols suitable for nursing practice.
Keywords: medical intensive care unit, mechanical ventilation, planned extubation, reintubation, risk prediction tool